Paul Krugman explains why one should believe in the Keynesian model. There are two main points he emphasizes:
First, we’re talking about a model, not just a prediction about the impact of spending increases. So you can ask about the ancillary predictions of that model as opposed to rival models. Anti-Keynesians assured us that budget deficits would send interest rates soaring; Keynesian analysis said they’d stay low as long as the economy remained far from full employment. Guess who was right?
Ricardian equivalence predicts that policy will not effect the interest rate.
The second point:
Also, there are some features of the approach that can be tested separately. Keynesianism isn’t just about sticky prices, but it does generally assume sticky prices — and there is overwhelming evidence, from a variety of sources, that prices are indeed sticky.
Evidence that prices do not adjust instantaneously does not imply that price stickiness is important in explaining fluctuations. For example, see this recent paper by Allen Head, Lucy Qian Liu, Guido Menzio, and Randy Wright. Here is the abstract:
Why do some sellers set prices in nominal terms that do not respond to changes in the aggregate price level? In many models, prices are sticky by assumption. Here it is a result. We use search theory, with two consequences: prices are set in dollars since money is the medium of exchange; and equilibrium implies a nondegenerate price distribution. When money increases, some sellers keep prices constant, earning less per unit but making it up on volume, so proﬁt is unaﬀected. The model is consistent with the micro data. But, in contrast with other sticky-price models, money is neutral.
Neither of these observations suggest that Krugman is wrong, but rather to note that there are other views that are observationally equivalent. In other words, his reasons for supporting Keynesian models is confirmation bias.